Freakonomics, Race and Groups

Steven Levitt, University of Chicago economist and co-author of Freakonomics, a new book co-written with journalist Stephen Dubner on Levitt's research, was a guest on the Daily Show last night. I was a bit intrigued by some of the findings he described (and also amused to hear someone trying to explain regression analysis on Comedy Central), so I looked into it a bit. One of the topics the book discusses, for example, is why there's such a gap in how blacks and whites perform on standardized testing. I found the article by Levitt on which the chapter is based (unfortunately, that link only works for UMN people as it's not available publicly). Here's the abstract:

This paper describes basic facts regarding the black-white test score gap over the first four years of school. Black children enter school substantially behind their white counterparts in reading and math, but including a small number of covariates erases the gap. Over the first four years of school, however, blacks lose substantial ground relative to other races; averaging .10 standard deviations per school year. By the end of third grade there is a large Black-White test score gap that cannot be explained by observable characteristics. Blacks are falling behind in virtually all categories of skills tested, except the most basic. None of the explanations we examine, including systematic differences in school quality across races, convincingly explain the divergent academic trajectory of Black students.

Levitt explains that there is good news in these findings, such as the fact that black students are disproportionately likely to come from poorer, less educated families/neighborhoods doesn't seem to have an effect on their abilities as they enter kindergarten. The bad news, though, is that something happens once they enter school: there's a steady increase in the gap each year after kindergarten. He also points out some other complications. For example,

In contrast to Blacks, Hispanics gain substantial ground relative to whites, despite the fact that they are plagued with many of the social problems that exist among blacks – low socioeconomic status, inferior schools, and so on.

So what's going on? They test a nice list of "leading hypotheses" on the issue, including:

the importance of parental and environmental contributions grow over time, black students suffer worse summer setbacks, standardized tests are poor measures, interactions between black students and schools interferes with learning, systematically lower quality schools, that differences may only manifest themselves in more involved, higher order problems.

They find none of these are good explanations and leave us scratching our heads. I, however, have a hypothesis that they didn't test and I think I'm right.

Ok, it's not really my hypothesis, I'm actually using Judith Harris' "group socialization theory" drawn from her book The Nurture Assumption (although the book is based on this journal article (free download) from the Psychological Review). Harris' controversial claim is that children aren't socialized by their parents, but by their peer groups. This flies in the face of our modern obsession with blaming all of our problems on bad child-parent experiences, but it makes a great deal of sense when you think about it. The way children act around their parents doesn't say much about how they act around their peers, and kids and adults view themselves as two completely different groups. As Harris puts it, children don't strive to be competent adults, they strive to be competent children. We can blame our parents for many things (our genes, whatever control they do have over which peer groups we end up associating with, etc.), but it's within peer groups that people are socialized.

Harris argues that people operate in completely different ways with personal relationships than they do with groups. This explains why saying "Some of my best friends are black" as a qualifier to a racist comment doesn't work. We think of personal relationships one way. Group relationships another. Parent-child relationships are personal relationships. Child-child relationships are group relationships.

Despite her use of research in behavioral genetics and evolutionary psychology to make her argument (two fields considered taboo by most sociologists - though not this one, I might add), her argument about group socialization is very sociological. In fact, much of what I wrote about groups in the previous post was partially inspired by the fact that I just completed her book. But her arguments about groups aren't totally new either, she just contextualizes them in a very unique way - by downplaying the effects of parents, emphasizing that genes effect personality and drawing on evolutionary psychology to help build her model of how people work. But sociologists/psychologists have long used the expression "stereotype threat" or "labeling theory" to explain how people internalize characteristics of the groups they are sorted (and sort themselves) into. Harris' contribution is by arguing for actual biological, psychological mechanisms to explain why these group processes happens as they do

So back to Levitt, et al. They look at lots of individual-level indicators, but, unless they don't mention it in their article, they don't look at "groupness" at all (hmmm, what a surprise for economists). One of their most interesting findings is that black students in white schools do as poorly as those in all black schools, and that the financial status of the school doesn't actually matter that much. They can't explain this, but they ignore one very basic, well-known stereotype in our culture: blacks aren't "supposed to be" the smart group. Whites think they're more dumb. Many blacks believe them (or reject white definitions of "smart"), and the charge that black students who do well in school are "acting white" is something everybody's heard about. According to "labeling theory," when black students are put into the category of "black," they will take on the stereotypical characteristics of that group. A self-fulfilling prophesy of sorts. This doesn't just happen with race. It happens with jocks, class clowns, boys and girls, geeks, nerds, preppies, hippies, goths, skaters, stoners, frat boys, sorority girls, jesus freaks and whatever other social clique you can imagine (and, of course, larger factors like the neighborhood culture, the media, etc. effect how the norms of these groups are defined).

Harris discusses this very topic and points out some interesting research. For example, black students who are children of black immigrants, not "African Americans," tend to not be accepted by the African American groups in their school and they tend to do much better in school. Additionally, in situations where the school is overwhelmingly white and there's only a few isolated black student, they are more likely to become parts of white peer groups and do well in school (or at least as well as the peer group they associate with). Harris also discusses research on black colleges and how, running contrary to the logic of desegregation, when everyone is black, then that no longer becomes the salient category around which groups are organized and the norms about what it means to "be black" in the classroom become relaxed. (The same is found in sex-segregated schools, by the way.)

Interestingly, much of the evidence about Hispanics they cite ties in perfectly with Harris' theory. For example, they point out that

Hispanics do not test particularly poorly on reading, even upon school entry. Controlling for whether or not English is spoken in the home does little to affect the initial gap or the trajectory of Hispanics.

They sound surprised, but this is exactly what many, many studies on such families in linguistics has found. Guess when the exceptions to this are? When kids live in "ethnic enclaves" where their peer groups are entirely speaking their native language. Move the parents, and their language and culture in the home, to a more "American" neighborhood, the kid will speak and act just like his peers, not the parents.

Additionally, and probably most importantly, this would explain why the effect seems to be caused by school: because that's when racial identity becomes salient as a component of group identity.

In other words, the group that you identify with matters for how you think of yourself and how you behave. This is not, I repeat not, a characteristic of black kids only. It's the way humans work. For example, I'm in graduate school with lots of "smart kids." If there's a talk or lecture that some of us should probably go to based on our own research interests, the odds of us going increase exponentially if we know our friends are going. If it's an activity that's okay with the rest of the group, count us in. If it's not, our odds of going on our own accord drop dramatically, regardless of whether or not we really should go based on our interests.

So how could Levitt test these additional hypotheses? Test to see if the group dynamics in a school/neighborhood matter. When looking at "mixed race" schools, does it matter how large the number of black students are? Are there ways to measure how salient the black/white divide is at schools? Perhaps at some schools, there are crosscutting groups (perhaps based on athletics, gender or any other non-racial social group imaginable) that minimize the salience of race as a group identity. At schools where there are too few to form a salient group of "black students," do those students perform differently?